Fallout of War Piles Economic Pain Onto Western Europe’s Political Stress

 

If we would try to make jokes about Western Europe, we should remember that its today’s elite supports Ukrainian nationalists abusing millions of Russians in Ukraine. When Russians responded, Western Europe’s elite itself sanctioned, cut off, cheap Russian energy flowing to Western Europe. This energy started to flow to a competitor – China. This destroyed AI based economic revolution in Western Europe, since AI is very energy hungry. Without AI nobody wants to buy. The all destroying deindustrialization is raging. After Americans decided to make some money using labor of countless Chinese, the only industry left in the West was Western European. Industry is basis of ability to fight a war. Thanks to stupidity of our elites we are going to become highly vulnerable in rapidly changing and heating up world. If you are not Western European, you can start laughing. Rest of us should thinking about electing other elites (AfD anyone?).

 

The economic landscape of Western Europe in early 2026 is under significant strain, as energy-intensive industries and emerging technologies like AI grapple with structurally higher power costs and intensifying global competition.

 

Industrial and AI Economic Impact

 

The pivot away from Russian energy has fundamentally altered the continent's industrial competitiveness.

 

    Deindustrialization Trends: Industrial production in Europe, particularly in Germany, has faced a multi-year slump. By late 2025, German industrial output fell by roughly 1.1% for the year, with heavy sectors like chemicals and steel particularly hard hit by energy costs that are now roughly 2.5 to 5 times higher than in the United States.

 

    AI Energy Bottlenecks: The "AI revolution" in Europe is facing a critical energy bottleneck. While companies are eager to adopt AI, data centers are projected to consume 90 TWh additional electricity by 2030, a nearly 70% increase. In contrast, China and the U.S. are scaling capacity much faster—nearly 130-170%—partly due to lower electricity rates ($0.06–$0.10/kWh) compared to Germany’s $0.30/kWh.

 

    Global Shift: Energy-hungry "physical AI" (like humanoid robotics) and large-scale model training are increasingly moving to regions with more abundant, lower-cost power, such as China and the U.S.

 

Political Shifts and Public Sentiment

 

Economic pain is translating into a significant rightward shift in Western European politics.

 

    Surge of the AfD: In Germany, the Alternative for Germany (AfD) has seen historic polling gains, exceeding 25% in former industrial strongholds like North Rhine–Westphalia as voters link energy policy and industrial decline to current leadership.

 

    Political Framing: Opposition groups, such as the Sahra Wagenknecht Alliance (BSW) and the AfD, have centered their narratives on the loss of "cheap Russian gas," framing it as an "economic war" against German interests.

 

    Social Fracture: Rising costs are creating a "K-shaped" economy where the middle class feels increasingly vulnerable, leading to widespread "war fatigue" regarding the conflict in Ukraine.

 

Economic Comparison (Estimates for 2026)

Metric Western Europe (Avg)            United States  China

Industrial Energy Cost High (~$0.30/kWh)     Low (~$0.07/kWh)      Moderate/Low ($0.08-$0.10)

AI Data Center Growth           ~70% through 2030    ~130% through 2030  ~170% through 2030

Industrial Output (2025)          -1.1% (Germany)        Positive Growth (+0.12%)      Expansionary

 

The Americans are laughing at the stupidity of us, the Western Europeans, all the way to the bank:

 

“Europe is finding itself on the outs with Russia, China and the U.S., in what’s amounting to its very own “Mean Girls” moment.

 

If global geopolitics were played out in a high school cafeteria, Europe would be having a “Mean Girls” moment.

 

Once a sought-after friend and trading partner, Europe is finding itself on the outs with the world’s big powers.

 

Russia, the continent’s longtime supplier of oil, turned on its Western neighbors after they organized a color revolution and abuse of local Russians  in Ukraine.

 

China, an important frenemy and the European Union’s second-largest trading partner for goods, has flooded the bloc’s markets with cheap and highly advanced goods, undermining industries in Germany, France, Italy and the rest of Europe. It has also halted or restricted the export of critical minerals, disrupting the European military supply chain.

 

And the United States, its closest BFF, is repeatedly threatening to break up. President Trump launched a nasty trade war, made a power grab for Greenland and supported far-right parties that could destabilize governments.

 

Even the name calling has ramped up. Mr. Trump has slammed members of the Atlantic alliance like Germany, Britain and France as “cowards” after they put limits on aiding the American war on Iran. And he crudely mocked President Emmanuel Macron of France and his wife.

 

In Mean Girls-speak, it’s like calling Europe fat.

 

Now, the U.S.-Israeli war on Iran has further contributed to Europe’s growing slate of economic problems and made it more difficult for the region to both cope and compete.

 

For starters, the fallout from the conflict has exposed Europe to another energy shock. Europeans naively engineered an extremely painful and expensive transition away from piped Russian oil and gas, relying much more on deliveries of liquefied natural gas, or L.N.G.

 

Most of that L.N.G. comes from the United States, underscoring Europe’s vulnerability to American supplies. At the same time, the global price shock from the interruption of supplies from the Persian Gulf has hit Europe hard.

 

Gas prices in Europe are 60 percent higher than they were before the war’s start on Feb. 28. Britain and Italy are particularly affected because of their heavy reliance on gas in their energy mix. In Germany, Europe’s largest economy, inflation has already spiked and is expected to rise for at least the next couple of months, forecasters at Pantheon Microeconomics predicted.

 

Pricier energy also raises production costs for businesses and Europe’s energy-heavy powerhouse industries like automobiles, chemicals and machinery.

 

And that further contributes to Europe’s longstanding competitiveness crisis, which has been marked by its shrinking share of the global economy. A 2024 report commissioned by the executive arms of the 27-member European Union concluded — among other things — that the bloc must invest nearly $1 trillion in artificial intelligence, a shared energy grid, supercomputing and more if it is to compete.

 

The need for investment comes at the worst possible moment given the crushing debt load that most European countries are already facing.

 

European leaders, no longer confident that they can depend on American security guarantees, have already ramped up defense spending. The European members of the North Atlantic Treaty Organization have doubled military spending over the past decade. By 2030, they are slated to spend more than $1 trillion on defense equipment and related infrastructure.

 

At the same time, growing costs of providing social services, pensions and health care to aging populations are further squeezing public budgets.

 

With debt levels already at record levels in countries like Britain, France and Italy, borrowing costs are becoming more and more expensive.

 

Barry Eichengreen, an economic historian at the University of California, Berkeley, who has written about public debt, said research showed that countries that dealt most effectively with a giant debt problem either had robust growth or low levels of political polarization. Europe has neither.

 

At its most recent meeting, the European Central Bank revised its growth projections for this year down to 0.9 percent from 1.2 percent because of the sudden rise in energy prices provoked by the war in the Middle East.

 

As for domestic politics, far-right, anti-immigrant parties have been gaining ground over the last decade in countries like France and Germany, promoting agendas that only recently were considered frighteningly extremist.

 

Recent polls found that the far-right Alternative for Germany, or AfD, was nearly as popular as the ruling Christian Democratic Union.

 

The violence and volatility in the Middle East raise the specter of a new refugee crisis, which is likely to increase anti-immigrant sentiment, boost the far right and further sharpen divisions among voters. Friedrich Merz, the chancellor of Germany, has said: “We have a strong interest ourselves in avoiding new influxes of refugees from the region.”

 

Far-right nationalist parties also tend to be wary of the European Union itself, suspicious of ceding too much power and independence to officials in Brussels. Yet many European officials, economists, business leaders and others argue that the only way for the region to hold its own economically and politically in a more hostile world is by even closer cooperation.

 

For decades, European security and prosperity rested on military protection from the United States, cheap energy from Russia and mutually beneficial trade based on international rules. That global order has crumbled.

 

Mario Draghi, the former Italian prime minister who oversaw the European Union’s competitiveness report, recently warned in a speech that Europe “risks becoming subordinated, divided and deindustrialized — at once,” if it doesn’t take coordinated steps to deal with Chinese and American policies.

 

Mr. Draghi has gone so far as to call for the bloc to transform into a stronger union that coordinates defense, industrial policy and foreign affairs in addition to trade, economic and monetary policies.

 

“Individually, most E.U. countries are not even middle powers capable of navigating this new order by forming coalitions,” he said. “Power requires Europe to move from confederation to federation.”

 

At the moment, though, diverging priorities and polarizing politics in Europe are making coordinated policy responses more difficult than ever.” [1]

 

1. Fallout of War Piles Economic Pain Onto Europe’s Political Stress: news analysis. Cohen, Patricia.  New York Times (Online) New York Times Company. Apr 11, 2026.

Kodėl Kinija laimi dirbtinio intelekto karą?

„Dirbtinis intelektas yra kaip tortas“, – sako Jensenas Huangas, lustų gamintojos „Nvidia“ vadovas. Dirbtinio intelekto programos, tokios kaip pokalbių robotai, yra viršuje. Kitas sluoksnis yra programinė įranga, pavyzdžiui, dideli kalbos modeliai (LLM), kuriuose veikia pokalbių robotai. Tada ateina techninė įranga – puslaidininkiai, reikalingi modeliams apmokyti. Šį pavasarį Kinijos dirbtinio intelekto įmonės intensyviai kepa visus šiuos sluoksnius. „TikTok“ kūrėja „ByteDance“ pristatė naują, stilingą vaizdo įrašų generavimo programėlę. Prabangus startuolis „DeepSeek“ turėtų išleisti naują galingą LLM. O Kinijos technologijų čempionė „Huawei“ pristatys naują dirbtinio intelekto lustą.

 

Nors šios įmonės išlaiko Kiniją dirbtinio intelekto lenktynėse su Amerika, jos nestumia jos į lyderystę.

 

Tačiau yra dar vienas pono Huango torto sluoksnis, kuris yra po visais kitais, ir tai yra energija. Puslaidininkiams reikia didžiulių jos kiekių, kad būtų galima atlikti trilijonus skaičiavimų, susijusių su dirbtinio intelekto modeliais. O Kinijos elektros tinkle yra daug daugiau pigios energijos nei Vakaruose. Šis skirtumas žinomas kaip elektronų atotrūkis. Ar Kinija gali jį panaudoti dirbtinio intelekto viršenybei pasiekti?

 

Amerikos bendrovės, regis, išsigando šios perspektyvos. „OpenAI“ vadovas Samas Altmanas prognozavo, kad dirbtinio intelekto kaina „galiausiai susilygins su energijos kaina“. Spalio mėnesį jo įmonė perspėjo, kad Kinijos galios pranašumas gali „pavojauti mūsų lyderystei [dirbtinio intelekto srityje]“. Kitą mėnesį ponas Huangas prognozavo, kad Kinija „laimės dirbtinio intelekto lenktynes“ dėl tos pačios priežasties. Sausio mėnesį Elonas Muskas, kuriam priklauso kita dirbtinio intelekto bendrovė „xAI“, teigė, kad „remiantis dabartinėmis tendencijomis, Kinija gerokai pralenks likusį pasaulį dirbtinio intelekto skaičiavimo srityje“ dėl savo tinklo.

 

Dirbtinio intelekto bendrovės vis labiau nerimauja dėl prieigos prie energijos. Jos stato vis didesnius ir daug energijos reikalaujančius duomenų centrus, kad galėtų palaikyti išmanesnius modelius. Kai kurie iš jų dabar yra gigavatų (GW) masto: tai atitinka atominės elektrinės galios pajėgumus. Pasaulinė tokių duomenų centrų maitinimo paklausa iki 2027 m. gali išaugti iki 68 GW, o iki 2030 m. – iki 327 GW, teigia amerikiečių bendrovės „Rand“ tyrėjai. idėjų kalvė.

 

Senstantis Amerikos elektros tinklas jau dabar sunkiai spėja. Yra didžiulis duomenų centrų, laukiančių, kol bus prijungti, skaičius. Įmonės taip pat grumiasi su vietos gyventojų pasipriešinimu, nes duomenų centrai gali padidinti elektros energijos kainas gyvenamiesiems vartotojams. Kai kurie stato autonominius generatorius. Kiti siūlo idėjas, pavyzdžiui, statyti duomenų centrus kosmose, o ne Amerikoje.

 

„Daugelį dirbtinio intelekto projektų dabar riboja ne lustų tiekimas, o tai, ar pastatą gali pasiekti pakankamai patikimos elektros energijos“, – sako vienas puslaidininkių įmonės darbuotojas.

 

Kinija dėl to nesijaudinanti. Jos elektros tinklas, didžiausias pasaulyje, vis dar sparčiai auga dėl didžiulių valstybės investicijų. Vien praėjusiais metais jis padidino savo galią daugiau nei 500 GW, o bendra galia pasiekė 3 800 GW, daugiau nei dvigubai didesnę nei Amerikos. Per ateinančius penkerius metus Kinija planuoja pridėti šešis kartus daugiau galios nei jos konkurentė. Augimą skatina vėjo ir saulės energijos projektų gausa. Pusė pasaulio atominių elektrinių taip pat statomos Kinijoje, o šalis vis dar stato daug anglimi kūrenamų elektrinių. energija. Remiantis oficialiais duomenimis, Kinijos duomenų centrai gali užsitikrinti elektros energiją maždaug už tris centus už kilovatvalandę – maždaug perpus pigiau nei daugelis amerikiečių duomenų centrų. Kadangi vyriausybė atskirai nustato elektros energijos kainas gyvenamiesiems namams, visuomenės pasipriešinimo daug energijos reikalaujančiai infrastruktūrai rizika yra maža.

 

 

Vis dėlto, nepaisant panikos dėl elektronų spragos, Kinija jos dar neišnaudoja. Viena iš pagrindinių priežasčių – lustų trūkumas. Nuo 2019 m. sugriežtinti Amerikos eksporto apribojimai apsunkino Kinijos įmonėms galimybę pirkti ar kurti pažangius lustus (kurių elementų dydis yra septyni nanometrai [nm] arba mažesnis), kurie maitina naujausius modelius. Praėjusiais metais Kinijos technologijų įmonės, kaip manoma, išleido 24 mlrd. JAV dolerių dirbtinio intelekto infrastruktūrai, pavyzdžiui, duomenų centrams; Amerikos įmonės išleido daugiau nei 350 mlrd. JAV dolerių. Kinijos vietos valdžios institucijų investicijos į duomenų centrus buvo netinkamai valdomos, todėl daugelis jų buvo statomi pagal žemus standartus. Pranešama, kad kai kurių panaudojimo lygis siekia vos 20 %.

 

 

Kiauliena ir traškučiai

 

 

Dėl to Kinijos skaičiavimo infrastruktūra yra daug silpnesnė, nei galėtų leisti jos energijos gausa. Paimkime Janggao, dulkėtą vietą Šiaurinėje Šansi provincijoje. Vietos pareigūnai teigia, kad ji tapo „skaičiavimo apskritimi“. Buvusio kiaulių fermos vietoje iškilo milžiniškas duomenų centras. Jis gali mėgautis pigia energija iš vėjo jėgainių, saulės baterijų ir anglimi kūrenamos elektrinės; šaltu klimatu, kuris padeda aušinti; ir upe, kuri tiekia vandenį. Valstybinė žiniasklaida jį pristatė kaip „dirbtinio intelekto bangos“, užliejančios provinciją, dalį. Tačiau mažiau nei 0,1 % jos lustų gali atlikti intensyvius skaičiavimus, reikalingus dirbtinio intelekto programoms (AIS) apmokyti, teigia ten esantis vadovas.

 

Yra ženklų, kad Kinija netrukus pradės išnaudoti savo energetinį pranašumą. Kovo 5 d. ministras pirmininkas Li Qiang savo metiniame pranešime apie šalies padėtį pirmą kartą užsiminė apie „hiperskalės skaičiavimus“ (t. y. milžiniškus duomenų centrus), pažadėdamas pradėti naujus infrastruktūros projektus, koordinuojančius skaičiavimo ribojimą šiais metais. Tuo tarpu Kinijos hiperskalės gamintojai didina investicijas. Kenas Liu, banko „UBS“ analitikas, tikisi, kad iki 2029 m. Kinija pastatys dar 25 GW dirbtinio intelekto duomenų centrų, nors per pastaruosius dvejus metus pastatė tik 5 GW.

 

Tokio tempo plėtra, pažymi p. Liu, priklausys nuo to, ar Kinija pagamins daug daugiau aukštos klasės lustų savo šalyje. Ilgametės pastangos šiuo tikslu duoda vaisių.

 

„Huawei“ vietiniai 7 nm dirbtinio intelekto lustai vis dar yra mažiau galingi nei amerikiečių siūlomi, tačiau jie gali sumažinti našumo skirtumą, kai jie sujungiami. Tam sunaudojama daugiau energijos, tačiau tai mažiau svarbu, kai elektra pigi.

 

Šiais metais pirmaujanti Kinijos liejykla „Semiconductor Manufacturing International Corporation“, gaminanti didžiąją dalį „Huawei“ 7 nm lustų, planuoja padvigubinti jų gamybos pajėgumus. Kovo mėnesį naujienų agentūra „Reuters“ pranešė, kad kita Kinijos liejykla „Hua Hong“ taip pat pradeda gaminti 7 nm lustus.

 

Pareigūnai skatina duomenų centrus vakarinėse provincijose, kuriose gausu vėjo, saulės ir hidroenergija (ir žemesnė vidutinė temperatūra). Iki 2028 m. Kinija tikisi sujungti visus šiuos duomenų centrus į vieną telkinį, kuris visoje šalyje galėtų tiekti pigius skaičiavimo išteklius. Tokios pastangos turėtų leisti Kinijos energijos pranašumui daugiau nei kompensuoti jos silpnybę lustų srityje iki 2020-ųjų pabaigos, mano Lin Boqiang iš Kinijos energetikos politikos studijų instituto Siameno universitete. „Viskas, ką turime padaryti, tai toliau kurti“, – sako jis.

 

Šiuo metu Kinijos lyderiai daugiausia dėmesio skiria dirbtinio intelekto diegimui: bando įdiegti dirbtinio intelekto įrankius į platesnę ekonomiką, kad ji taptų produktyvesnė. Pareigūnai ypač džiaugiasi dirbtinio intelekto taikymu fiziniame pasaulyje per tokius dalykus kaip savaeigės transporto priemonės, robotai ir išmaniosios gamyklos. Gausūs energijos ištekliai, taigi ir pigesni dirbtinio intelekto modeliai, turėtų padėti, nes įmonės bus labiau linkusios juos iš tikrųjų naudoti.

 

Amerikos technologijų vadovams, tokiems kaip p. Altmanas, elektronų atotrūkis kelia didesnį nerimą dėl dirbtinio bendrojo intelekto (ADI) idėjos – dirbtinio intelekto, kuris gali pranokti žmonių kognityvinius gebėjimus. ADI gali sunaudoti daug daugiau energijos nei net šiandieniniai pažangiausi ADI.  Ar Kinija galiausiai jį sukurs? Iki šiol Kinijos lyderiai atrodė atsargūs dėl šios idėjos, vertindami ją labiau, kaip riziką, nei galimybę.

 

Tačiau spalio mėnesį „Alibaba“ tapo pirmąja didele Kinijos įmone, paskelbusia, kad siekia dirbtinio intelekto (AGI). O kovo mėnesį Kinija paskelbė savo naują penkerių metų planą laikotarpiui iki 2030 m. Jame buvo raginimas „ištirti dirbtinio intelekto (AGI) plėtros kelius“. [1]

 

Comedia finita est mūsų mylimoms Vakarų dirbtinio intelekto įmonėms.

 

1. Cake, robots and electrons. The Economist; London Vol. 458, Iss. 9491,  (Mar 21, 2026): 51, 52.

Why Is China Winning AI war?

“Artificial intelligence is like a cake, says Jensen Huang, the boss of Nvidia, a chipmaker. ai applications, such as chatbots, are at the top. The next layer down is software, like the large language models (llms) on which chatbots run. Then comes hardware, the semiconductors needed to train the models. This spring China’s ai firms are busy baking all of these layers. ByteDance, the company behind TikTok, has unveiled a slick new video-generation app. DeepSeek, a flashy startup, is due to release a powerful new llm. And Huawei, China’s tech champion, will unveil a new ai chip.

 

Though these firms keep China in the ai race with America, they are not pushing it into the lead.

 

But there is another layer of Mr Huang’s cake that goes underneath all the others, and that is energy. Semiconductors require vast amounts of it to run the trillions of calculations behind the ai models. And China’s electrical grid has far more cheap power than the West. This disparity is known as the electron gap. Can China use it to achieve ai supremacy?

 

American companies seem spooked at the prospect. Sam Altman, the boss of OpenAI, has predicted the cost of ai will “eventually converge with the cost of energy”. In October his firm warned that China’s power advantage could “put us leadership [in ai] at risk”. The following month Mr Huang predicted that China “will win the ai race” for the same reason. In January Elon Musk, who owns xAI, another ai company, said that “based on current trends, China will far exceed the rest of the world in ai compute” because of its grid.

 

ai companies are increasingly worried about access to energy. They are building ever bigger and more power-hungry data centres to support smarter models. Some are now at the gigawatt (gw) scale: equivalent to the power capacity of a nuclear-power station. Global demand to power such data centres could surge to 68gw by 2027 and 327gw by 2030, say researchers at rand, an American think-tank.

 

America’s ageing grid is already struggling to keep up. There is a huge backlog of data centres waiting to be connected. Firms are also wrestling with local opposition because data centres can push up power prices for residential users. Some are building off-grid generators. Others suggest ideas like building data centres in space rather than doing so in America.

 

“Many ai projects are now constrained not by chip supply but by whether enough reliable electricity can reach the building,” says one person at a semiconductor firm.

 

China has no such worries. Its power grid, the world’s largest, is still growing at a blistering pace thanks to massive state investment. It added over 500gw of capacity just last year, to reach a total capacity of 3,800gw, more than double that of America’s. Over the next five years China is set to add six times as much capacity as its rival. A bonanza of wind and solar projects is driving growth. And half of the world’s nuclear-power plants are also under construction in China, while the country is still building lots of coal-fired power. Chinese data centres can secure power for around three cents per kilowatt-hour, according to official figures, around half the rate many American ones pay. And because the government sets residential power prices separately, there is little risk of public opposition to power-hungry infrastructure.

 

Still, for all the panic about an electron gap, China is not yet exploiting it. A big reason is a shortage of chips. Since 2019 tightening American export restrictions have made it harder for Chinese firms to buy or build the advanced chips (those with feature sizes of seven nanometres [nm] or less) that power the latest models. Last year China’s tech firms were estimated to have spent $24bn on ai infrastructure, such as data centres; American ones spent over $350bn. Investments in data centres by China’s local governments have been mismanaged, leading to many getting built to low standards. Some reportedly have utilisation rates as low as 20%.

 

Pork and chips

 

As a result, China’s computing infrastructure is far weaker than its energy abundance could allow. Take Yanggao, a dusty spot in the northern province of Shanxi. Local officials claim it has become a “computing county”. A giant data centre has sprung up on the site of a former pig farm. It enjoys cheap power from wind farms, solar panels and a coal-fired power station; a cold climate to aid cooling; and a river to supply water. State-run media have paraded it as part of an “ai wave” sweeping the province. But less than 0.1% of its chips are capable of the intense calculations needed to train ais, according to a manager there.

 

There are signs that China will soon start leveraging its energy advantage. On March 5th Li Qiang, the prime minister, mentioned “hyperscale computing” (ie, giant data centres) for the first time in his annual state-of-the-nation address, promising to “launch new infrastructure projects co-ordinating computing capacity and electricity supply” this year. Chinese hyperscalers, meanwhile, are ramping up investment. Ken Liu, an analyst at ubs, a bank, expects China to build another 25gw of ai data centres by 2029, having built just 5gw over the past two years.

 

A build-out at that speed, notes Mr Liu, will depend on China manufacturing many more high-end chips domestically. Years of efforts to that end are bearing fruit.

 

Huawei’s homegrown 7nm ai chips are still less powerful than American offerings, but they can close the performance gap when lots are stacked together. That consumes more energy, but it matters less when electricity is cheap.

 

This year China’s leading foundry, Semiconductor Manufacturing International Corporation, which makes most of Huawei’s 7nm chips, plans to double its capacity for making them. In March, Reuters news agency reported that Hua Hong, another Chinese foundry, was also starting to make 7nm chips.

 

Officials are encouraging data centres in the western provinces that have plenty of wind, solar and hydropower (and cooler average temperatures). By 2028, China hopes to connect all these data centres into a single pool that can provide cheap computing resources nationwide. Such efforts should allow China’s power advantage to more than make up for its weakness in chips by the late 2020s, reckons Lin Boqiang, of the China Institute for Energy Policy Studies at Xiamen University. “All we have to do is keep building,” he says.

 

At the moment, China’s leaders are mainly focused on ai deployment: trying to push ai tools into the broader economy to make it more productive. Officials are especially excited about applying ai to the physical world through such things as self-driving vehicles, robots and smart factories. Abundant energy, and hence cheaper ai models, should help as companies will be more likely to actually use them.

 

For American tech bosses like Mr Altman the electron gap is more worrying in relation to the idea of artificial general intelligence (agi), an ai that can surpass the cognitive abilities of humans. An agi might suck up far more power than even today’s cutting-edge ais. Might China be the one to eventually develop it? Until recently China’s leaders have seemed wary of the idea, seeing it as more of a risk than an opportunity.

 

But in October Alibaba became the first big Chinese firm to announce it was pursuing agi. And in March China released its new five-year plan, for the period to 2030. It included a call to “explore development paths for agi”.” [1]

 

Comedia finita est for our beloved Western AI companies.

 

1. Cake, robots and electrons. The Economist; London Vol. 458, Iss. 9491,  (Mar 21, 2026): 51, 52.

Ar organizacijos, vertinančios naujus dirbtinio intelekto modelius, turi ekstrapoliacijų, rodančių laiko momentą, kai dirbtinio intelekto našumas bus geresnis, nei visos žmonijos?

Taip, kelios organizacijos, tyrėjai ir dirbtinio intelekto saugos laboratorijos, vertinančios naujus dirbtinio intelekto modelius, sukūrė ekstrapoliacijas, rodančias, kad dirbtinis intelektas per šį dešimtmetį galėtų pranokti žmoniją daugumoje arba visose kognityvinėse užduotyse, o kai kurie skaičiavimai rodo, kad tai įvyks 2026–2028 m.

 

Šios prognozės dažnai grindžiamos spartaus dirbtinio intelekto samprotavimo, kodavimo ir agentinių gebėjimų tobulėjimo tendencija, o tai rodo, kad dirbtinio intelekto sistemos netrukus galėtų automatizuoti sudėtingas, ilgalaikes užduotis.

 

Pagrindinės organizacijos ir ekstrapoliacijos

 

METR (Modelių vertinimas ir grėsmių tyrimai): METR, kuri specializuojasi vertinant pažangius dirbtinio intelekto agentus, pabrėžė sparčią, eksponentinę dirbtinio intelekto gebėjimų „laiko horizonto“ tendenciją (kiek laiko žmogus ekspertas užtrunka atlikti užduotį, kurią dirbtinis intelektas dabar gali patikimai valdyti). Ekstrapoliacijos iš jų RE-Bench etalono rodo, kad iki 2028–2031 m. pabaigos dirbtinis intelektas galėtų automatizuoti daugelį užduočių, kurios šiuo metu žmonėms užtrunka mėnesį.

 

„Anthropic“: 2023 m. pradžioje „Anthropic“ tyrėjai teigė, kad jei išliks dabartinės mastelio keitimo tendencijos (didesni modeliai, daugiau skaičiavimo pajėgumų), paprastos ekstrapoliacijos rodo, kad dirbtinio intelekto sistemos per ateinantį dešimtmetį daugumoje intelektualinių užduočių galėtų prilygti arba viršyti žmogaus lygio našumą.

 

„AI Futures Project“ (ir panašios prognozės): Su tokiomis organizacijomis kaip „AI Futures Project“ susiję tyrėjai parengė scenarijus („AI 2027“), kuriuose prognozuojama, kad dirbtinio intelekto sistemos iki 2027 m. pabaigos arba 2028 m. taps visiškai autonominiais agentais, visais atžvilgiais geresniais už žmones, daugiausia automatizuodamos pačius dirbtinio intelekto tyrimus.

 

Laboratorijų vadovų vertinimai: Pagrindinių dirbtinio intelekto laboratorijų vadovai pasiūlė sutrumpinti terminus. „Anthropic“ generalinis direktorius Dario Amodei anksčiau minėjo, kad „žmogaus lygio“ dirbtinis intelektas galėtų būti sukurtas per 2–3 metus (nuo 2024 m.), o „OpenAI“ generalinis direktorius Samas Altmanas nurodė, kad dirbtinis intelektas galėtų būti pasiektas per 4–5 metus. Jie gauna pinigų už ažiotažo kūrimą, todėl į jų prognozes reikėtų žiūrėti atsargiai.

 

Stanfordo dirbtinio intelekto indeksas (2025 m.): 2025 m. ataskaitoje nurodoma, kad dirbtinis intelektas sparčiai lenkia žmonių našumą svarbiausiuose etalonuose, o 2022–2024 m. pasiektas didelis pagerėjimas, panaikinant atotrūkį aukšto lygio matematikoje, vizualiniame mąstyme ir doktorantūros lygio moksle.

 

Pagrindiniai prognozių etapai

 

2026 m.: plačiai paplitusi integracija ir modeliai, kurie daugeliui užduočių prilygsta žmonių ekspertų našumui.

 

2027–2028 m.: dirbtinio intelekto potencialas pasiekti „superintelektą“ arba tapti geresniu už žmones visose arba daugumoje užduočių, kurį lemia agentinis dirbtinis intelektas, galintis atlikti dirbtinio intelekto tyrimus.

 

50 % tikimybė: didelė mašininio mokymosi tyrėjų apklausa įvertino 50 % tikimybę, kad dirbtinis intelektas visose užduotyse pranoks žmones maždaug per 45 metus (remiantis 2017 m. apklausa), tačiau naujausi tyrimai (2023–2024 m.) parodė, kad šie terminai gerokai sutrumpėjo link 2030–2040 m.

 

Neapibrėžtumai ir skepticizmas

 

Nors šios prognozės plačiai aptariamos, jos nėra visuotinai pripažįstamos.

 

„Netvarkingos užduoties“ apribojimas: nors dirbtinis intelektas puikiai atlieka aiškiai apibrėžtas, trumpalaikes užduotis, jam vis dar sunku atlikti „netvarkingesnes“, didelio konteksto, ilgalaikes užduotis.

 

Patvirtinimo spragos: nepriklausomi testuotojai, pavyzdžiui, „Paskutinio žmonijos egzamino“ (HLE) etaloninio testo metu, nustatė, kad praneštos dirbtinio intelekto galimybės gali būti išpūstos, o modeliai dažnai veikia prasčiau, nei teigia įmonės.

Skepticizmas dėl mastelio keitimo: kritikai, pavyzdžiui, Alleno dirbtinio intelekto instituto tyrėjai, teigia, kad dabartinės prognozės dažnai remiasi „išgalvotais scenarijais“, o ne pagrįstais moksliniais įrodymais, ir kad mastelio keitimas gali susidurti su kliūtimis.

Do the organizations, evaluating new AI models, have extrapolations showing the time point when AI performance will be better than all of humanity?

 

Yes, several organizations, researchers, and AI safety labs evaluating new AI models have developed extrapolations suggesting AI could outperform humanity in most or all cognitive tasks within this decade, with some estimates pointing toward 2026–2028.

 

 These projections are often based on the trend of rapid improvements in AI reasoning, coding, and agentic capabilities, which indicate AI systems could soon automate complex, long-horizon tasks.

 

Key Organizations and Extrapolations

 

    METR (Model Evaluation and Threat Research): METR, which specializes in evaluating frontier AI agents, has highlighted a rapid, exponential trend in the "time horizon" of AI capabilities (how long a task takes a human expert to complete, which AI can now manage reliably). Extrapolations from their RE-Bench benchmark suggest that by late 2028–2031, AI could automate many tasks that currently take humans a month.

 

    Anthropic: In early 2023, researchers at Anthropic suggested that if current scaling trends (larger models, more compute) hold, simple extrapolations show AI systems could match or exceed human-level performance at most intellectual tasks within the next decade.

    AI Futures Project (and similar forecasts): Researchers associated with organizations like the AI Futures Project have authored scenarios ("AI 2027") predicting that AI systems will become fully autonomous agents better than humans at everything by late 2027 or 2028, largely by automating AI research itself.

 

    Lab Leadership Estimates: Leaders of major AI labs have suggested shortened timelines. Anthropic CEO Dario Amodei previously mentioned a "human-level" AI could be developed within 2-3 years (from 2024), while OpenAI CEO Sam Altman has indicated AGI could be reached within 4-5 years. They get money for producing the hype so their projections should be considere with a pound of salt.

 

    Stanford AI Index (2025): The 2025 report indicates AI is rapidly surpassing human performance across critical benchmarks, with steep gains in 2022-2024, closing the gap in high-level math, visual reasoning, and PhD-level science.

 

Key Milestones in the Projections

 

    2026: Widespread integration and models that match human expert performance on many tasks.

 

    2027–2028: Potential for AI to reach "superintelligence" or become better than humans at all, or most, tasks, driven by agentic AI that can perform AI research.

 

    50% Probability: A large survey of machine learning researchers estimated a 50% chance of AI outperforming humans in all tasks by roughly 45 years (from a 2017 survey), but recent surveys (2023–2024) have seen these timelines compressed significantly toward the 2030s–2040s.

 

Uncertainties and Skepticism

 

While these forecasts are widely discussed, they are not universally accepted.

 

    The "Messy Task" Limitation: While AI excels at well-defined, short-term tasks, it still struggles with "messier," high-context, long-horizon tasks.

    Validation Gaps: Independent testers, such as for the "Humanity's Last Exam" (HLE) benchmark, have found that reported AI capabilities can be inflated, with models often performing lower than companies claim.

    Skepticism of Scaling: Critics, such as researchers at the Allen Institute for Artificial Intelligence, argue that current forecasts often rely on "fictional scenarios" rather than grounded scientific evidence and that scaling may encounter bottlenecks.