Semiconductor behemoth Nvidia announced second-quarter sales doubled to $96.2 billion, temporarily reassuring US tech investors after a summer of mounting market anxiety over artificial intelligence spending.
Nvidia chief executive Jensen Huang, who is routinely seen wearing a leather jacket, confirmed that artificial intelligence has reached a commercial turning point where processing power directly generates corporate revenue.
Huang said: "AI has reached its inflection point. It’s doing useful work. Its tokens - units of data - are productive and profitable. Now, compute is revenue."
Despite the stellar financial results from the world's most valuable company, which serves as a benchmark proxy for the wider technology sector, a new spectre has emerged to threaten market celebrations. Investors are increasingly turning to the debasement trade, moving capital into gold bullion and bitcoin as precious metals and cryptocurrencies look more alluring than technology stocks or general equities.
The rush into bullion and bitcoin was sparked by Donald Trump's stand-off with trading partners of Iran, including China, alongside a controversial bond market intervention by US Treasury Secretary Scott Bessent. Wall Street hedge fund manager Stanley Druckenmiller, who mentored Bessent, has described the Treasury secretary's move as hazardous.
At the same time, rising borrowing costs pose a direct threat to corporate technology balance sheets. David Coombs of wealth management firm Rathbones pointed out that US bond yields rising above 5 per cent will spell bad news for Big Tech companies, which have accumulated prodigiously high borrowings to fund the race for artificial intelligence dominance.
Coombs described current conditions as "bewildering times" for tech investors, while noting that individuals cannot turn their back on artificial intelligence because "it’s real and growing exponentially, and changing lives." Investors who put £1,000 into each of the Magnificent Seven technology giants - Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla - four years ago would have more than tripled their money, but market experts say past strategies may no longer suffice.
Rising debt and data centre opposition
The rapid ascent of artificial intelligence is widely compared to the industrial revolution, yet growing worries over financial returns on capital expenditure threaten to weigh on financial markets in the months ahead. Nvidia's record earnings failed to dispel widespread uncertainty surrounding the eventual payback from $4.2 trillion in total AI borrowing and spending planned by major hyperscalers.
Hyperscalers - the large technology groups building and running the massive data centres that power artificial intelligence applications - include Google parent company Alphabet, Amazon, Apple, Meta, Microsoft and Oracle. All six corporations are major customers of Nvidia, whose microchips are regarded as the computational workhorses of AI.
To ensure customer loyalty, Nvidia helped the hyperscalers secure funding through a $500 billion financing partnership with Wall Street institutions including Apollo and Goldman Sachs. Although Nvidia denied that the arrangement represents risky circular financing, the multi-billion-dollar deal has ratcheted up broader market anxiety.
Data centre expansion is also confronting intense local community opposition. These facilities, filled with server racks and dense cabling, are voracious consumers of electricity and water. American residents are increasingly adopting Nimby - Not In My Back Yard - stances against local data centre construction, a movement analysts warn could spread to other nations.
Wall Street ratings and corporate legal risks
Despite mounting financial and operational challenges, Wall Street analysts remain surprisingly sanguine about Big Tech prospects, maintaining "buy" ratings on Alphabet, Amazon, Apple, Microsoft, Meta and Oracle. Nvidia shares also remain rated as a "buy", even though its chips face competition from low-cost Chinese rivals and custom processors being developed internally by Alphabet and Amazon. Investors who put £10,000 into Nvidia four years ago would hold £110,000 today.
Moral concerns continue to deter some investors from Meta, the corporate parent of WhatsApp and Instagram, despite the company settling a major US court case over claims its platforms were harmfully addictive to children and teenagers. Matt Britzman, an analyst at brokerage Hargreaves Lansdown, argued that the $1.5 trillion corporation's share price already reflects potential legal liabilities and its prolific AI spending.
Investor optimism is also being buoyed by expectations surrounding the forthcoming initial public offering of Anthropic, creator of the Claude AI system. Anthropic could be valued at $2 trillion on the basis that the total addressable market for its artificial intelligence services could eventually reach $30 trillion.
That $30 trillion market claim exceeds the $28.5 trillion figure cited during the blockbuster June initial public offering of Elon Musk's rocket and satellite company SpaceX. While it remains unclear whether Anthropic will allow small retail investors to participate in the flotation as SpaceX did, retail investors can access indirect exposure through investment trusts holding stakes in the firm, including RIT Capital Partners, Schiehallion, Baillie Gifford US Growth, and Scottish Mortgage.
Strategies for diversifying AI investment portfolios
Rather than relying solely on mega-cap tech stocks, Coombs advised that investors back businesses across the AI value chain by purchasing shares in companies involved in every stage of technological development. His recommended list of American companies includes Alphabet, Amazon, semiconductor maker AMD, chip design software group Cadence, data-centre operator Equinix, and energy company WEC, whose expertise in renewable energy is becoming increasingly essential as electricity demand surges.
Investment bank Saxo Bank highlighted four additional US companies benefiting from data centre construction: energy supplier Dominion Energy, energy equipment manufacturer GE Vernova, heavy construction equipment maker Caterpillar, and machinery titan Deere. Although Deere is best known for manufacturing agricultural tractors, its bulldozers and excavators are heavily utilized to clear construction sites for new data centres.
For investors seeking international diversification outside the United States, Coombs pointed to ASML, the Dutch manufacturer of specialized chip-making equipment, alongside Taiwanese electronics manufacturer Delta Electronics and Taiwan Semiconductor Manufacturing Company, known as TSMC.
Alternative routes into the sector include private equity allocations and specialized funds, such as direct stakes in SpaceX, the Polar Capital Technology AI Maximalist investment trust, and Allianz Technology. Market commentators consider these holdings essential components of a modern portfolio, though they warn that technology investments are increasingly among the riskiest assets an investor can hold.
