人工智能熱潮正在推動經濟發展。如果它失敗會怎樣？ （3/3）

Recently the New York Times reported the following:

The A.I. Boom Is Driving the Economy. What Happens if It Falters? (3/3)

A windfall for companies that build data centers and their suppliers is overshadowing weakness in other industries.

By Ben Casselman and Sydney Ember

(continue)

Uneven Effects

That abundance isn’t reaching everyone, however, or everywhere. The companies developing leading A.I. models, like OpenAI and Google, are mostly headquartered in the San Francisco area, and their highest paid employees are concentrated in the large “superstar” cities that have been the winners in the 21st century economy.

Investments in A.I. infrastructure are far more spread-out, but the benefits are much less certain. Some of the largest data centers are being built in rural areas, where land is cheap and jobs have been disappearing for decades. Local leaders in many communities have embraced data centers and other A.I. infrastructure as a way to diversify their economies and get a toehold in a burgeoning industry.

But data centers voraciously consume water and electricity, which can drive up costs for residents. And while they create jobs during construction, they employ relatively few people once they are operational, turning the sprawling campuses into humming ghost towns. Microsoft said, for instance, that its data centers employ about 50 people per building over a 24-hour period.

Many big data-center companies, known as “hyperscalers,” have won generous tax incentives from local communities, limiting their benefit to residents. And while local leaders argue the projects will attract other technology companies, it isn’t clear that will actually happen, said Mr. Muro, the Brookings economist.

“Some places have gotten confused by a lot of messaging from the hyperscalers that this is the beginning of a regional technology economy,” he said. Instead, he added, “These become just massive buildings with a couple hundred jobs, which aren’t terrible, but aren’t going to really move the dial.”

Still, many economists and industry experts say the infrastructure-building phase of the A.I. boom has room to run. Even as data center capacity expands at a record pace, demand is growing just as quickly, and companies are reporting yearslong backlogs. The demand for infrastructure is likely to keep growing, said Paul Ashworth, the chief North America economist at Capital Economics.

“It’s really only just beginning to get going,” he said. “The stock market may be close to a bubble,” he added, but there is little sign that the industry has bought too many chips or built too many data centers.

For these investments to pay off, however, artificial intelligence will need to fulfill its promise not just as a useful tool, but as a transformational technology that leads to huge increases in productivity.

If that doesn’t happen?

“A lot of the investment that has been put in place might turn out to be unjustified,” said Mr. Bhave, the Bank of America economist.

Translation

人工智能熱潮正在推動經濟發展。如果它失敗會怎樣？ （3/3）

數據中心建設公司及其供應商獲得的意外之財掩蓋了其他行業的疲軟。

（繼續）

影響不平均

然而，這種繁榮並沒有惠及所有人或所有地方。開發領先人工智能模式的公司，例如 OpenAI 和 Google，大多總部設在舊金山地區，其薪資最高的員工也集中在那些在 21 世紀經濟中脫穎而出的「超級明星」城市。

對人工智能基礎設施的投資分佈更為分散，但收益卻遠沒有那麼確定。一些最大的數據中心建在農村地區，那裡的土地價格低廉，幾十年來就業機會一直在消失。許多社區的地方領導人已將數據中心和其他人工智能基礎設施視為實現經濟多元化, 並在蓬勃發展的行業中佔據一席位的一種方法。

但數據中心耗水耗電巨大，會推高居民的生活成本。雖然數據中心在建設期間能創造就業機會，但一旦投入運營，員工人數卻相對較少，龐大的園區最終會變成空蕩蕩的「鬼城」。例如，微軟表示，其數據中心每棟大樓每24小時僅僱用約50名員工。

許多大型數據中心公司，即所謂的“超大規模雲端服務商”，都從當地社區獲得了豐厚的稅收優惠，這使得居民受益有限。布魯金斯學會經濟學家Muro先生表示，儘管當地領導人認為這些計劃將吸引其他科技公司，但這種情況是否真的會發生尚不清楚。

他說:「有些地方被超大規模雲端服務商的宣傳誤導，以為這是區域科技經濟的開端」。他並反而說: 「這些最終只會變成擁有幾百個工作崗位的大型建築，這雖然不算糟糕，但並不會真正帶來任何改變」。

儘管如此，許多經濟學家和產業專家仍然認為，人工智能熱潮的基礎建設階段仍有發展空間。儘管數據中心容量正以創紀錄的速度擴張，但需求也以同樣的速度成長，企業紛紛表示訂單積壓已達數年之久。Capital Economics北美首席經濟學Paul Ashworth表示，對基礎設施的需求可能會持續成長。

他說:「這實際上才剛剛開始」。 他補充道:「股市可能接近泡沫」，但幾乎沒有跡象表明該行業已經購買了過多的晶片或建造了過多的數據中心。

然而，要使這些投資獲得回報，人工智能不僅需要兌現其作為有用工具的承諾，更需要兌現其作為變革性技術的承諾，從而大幅提高生產力。

如果這種情況沒有發生呢？

美國銀行經濟學家Bhave先生說:「許多已經投入的資金, 最終可能會被證明是不值得的」。

              So, the U.S. economy in 2025 is split into two: Everything tied to artificial intelligence is booming. Just about everything else is not. It’s a two-track economy. This A.I. investments are rippling outward, propping up pieces of the economy outside the traditional technology sector. For example, Caterpillar, typically associated with AI with its bulldozers, Johnson Controls which is associated with cooling and fire-suppression systems, and Eaton Corp., which makes power-management systems. That abundance isn’t reaching everyone. For the AI investment boom to pay off, however, artificial intelligence will need to fulfill its promise not just as a useful tool, but as a transformational technology that leads to huge increases in productivity. Apparently, all these only time will tell.