AI Infrastructure Overtakes Crypto In The Race For Capital

The slowdown of the crypto market does not indicate a drying up of capital. It reveals their redeployment. They now fuel another investment cycle. While bitcoin struggles to move away from 63,037 dollars, the giants of Tech and finance concentrate their capital on infrastructures related to artificial intelligence and energy, to the detriment of cryptos.

Bear market money turns to real estate

Amidst the slump in crypto prices, the former Binance CEO, Changpeng Zhao, highlighted the abundance of capital available outside the crypto sector. The founder of the world’s leading exchange platform declared unequivocally on social media : “we may be in a bear market period, but there is a tremendous amount of capital seeking investment opportunities”.

Faced with this mass of liquidity seeking returns, Social Capital founder Chamath Palihapitiya gave a clear answer about the optimal destination of these funds, stating that hardware AI components should now be ignored in favor of land, power supply, and bare buildings. According to the American investor, this infrastructural triptych represents the most effective lever of the modern market, summarized by his formula : “the LPS triptych (Land, Power, Structure) remains the most obvious and fastest way to generate a return on equity”.

The market analysis rationally explains this strategic shift away from semiconductors and virtual tokens, based on several decisive factors :

• The dead end of chip-focused startup model: having co-founded Groq in 2016, whose technology was the subject of a non-exclusive license agreement with Nvidia in December 2025 valued by Palihapitiya at 20 billion dollars, the investor judges the sector increasingly difficult for new entrants ;
• Major industrial hurdles : the extreme constraints linked to manufacturing pace, the surgical precision required by factories, and chronic RAM shortages limit hardware component profitability ;
• The shift of scarcity towards energy : it is now direct access to electrical grids and exploitable land that represents the true bottleneck of the technological revolution ;
• A massive accumulation strategy : this is why Palihapitiya and his partner Anita Vlallian have accumulated nearly 6 gigawatts of deliverable energy capacity by 2029, betting that direct ownership of power supply surpasses mere computing power race.

The paralysis of Data Centers and the pivot of bitcoin mining actors

This rush towards electric real estate is explained by an unprecedented regulatory and social blockage on American soil. According to data published by Data Center Watch, over 75 data center projects amounting to approximately 130 billion dollars in investments were stopped or delayed during the first quarter of this year, a historical record. Local opposition groups have doubled their presence to cover 49 States, while legislators introduced more than 300 bills relating to data centers in six weeks, with the State of Maine missing by a single vote the outright ban of these facilities in the House.

In this saturation context, former bitcoin mining players hold a decisive competitive advantage thanks to their infrastructures already connected to the grid. This is the case of TeraWulf, listed on Nasdaq, which signed in July 2026 a 20-year lease estimated at 19 billion dollars with AI giant Anthropic to rent its 401 megawatt site located in Hawesville, Kentucky.

This massive transition of mining infrastructures towards artificial intelligence triggers a heated technical debate within the financial community. Reacting to these large maneuvers, Coinbase CEO Brian Armstrong sought to clarify the direct consequences for the Bitcoin network by reminding that “the hashing power or energy dedicated to bitcoin mining does not determine its price (the network difficulty adjusts if specialists disconnect to maintain the same block creation pace)”.

On his side, Jordi Visser from 22 V Research brings a note of caution and indicates that the era of easy capital in AI is ending, now forecasting normalized returns of around 30% per year. Moreover, the valuation of long-term leases like TeraWulf’s depends on future revenue projections that remain contingent on the delivery of energy, scheduled between the second half of 2027 and early 2028.

Repercussions on the crypto ecosystem

In the long term, this redeployment of bear market capital towards heavy infrastructure could redefine the overall economic balance between artificial intelligence and blockchain.

If the temporary drying up of liquidity on crypto markets in favor of energy penalizes bitcoin’s short-term price, the partial conversion of mining companies into highly profitable power lessors allows them to strengthen their cash flow amid market uncertainties.

However, the sustainability of this model remains subject to maintaining land scarcity and the ability of players to honor their contractual commitments by 2028. The key question for upcoming quarters lies in the possibility of seeing these colossal profits reinjected later into the crypto ecosystem, thus marking the return of liquidity to its original market once tensions on AI infrastructures stabilize.