The Phantom Commitment and the Quiet Delete 

The Phantom Commitment and the Quiet Delete 

On the evening of February 9, 2026, the White House published a fact sheet. It stated, in clear language, that India had agreed to import “certain pulses” from the United States. It stated that India was “committed” to purchasing more than $500 billion of American energy, agricultural goods, and technology products. It stated that India would “remove its digital services tax” . 

By the following afternoon, all of it was gone. Not through a press conference, not through a joint statement issued in New Delhi, not through a diplomatic clarification. The words simply vanished from the White House website, replaced with softer synonyms. “Commits” became “intends.” Pulses disappeared from the agricultural list altogether. The digital services tax removal—a long-standing American ask—was erased without explanation . 

This is the strange anatomy of the India-US interim trade framework: a deal that was written, then partially unwritten, in less than twenty-four hours. To read the original text and its revision side by side is to witness, in real time, the gap between what Washington wanted to claim and what New Delhi was actually willing to sign. The pulses were never in the joint statement issued by Prime Minister Narendra Modi and President Donald Trump. The $500 billion figure was never ratified by the Indian cabinet. The digital tax commitment was never cleared by the finance ministry. Someone in the White House—whether overeager staff or deliberate negotiator—tried to cement these concessions into the public record before India could object. India objected. The words disappeared . 

This is not, however, a story of American overreach corrected by Indian pushback. It is a story of how trade deals are now made between two capitals that do not entirely trust each other, that communicate through document edits and anonymous leaks, and that have learned to treat every published line as provisional until proven otherwise. The interim framework may have lowered tariffs from 50 percent to 18 percent, but the real negotiation—over what India actually promised, and what the United States actually secured—is still being fought, one deleted paragraph at a time . 

GPUs, Data Centres, and the $100 Billion Silicon Gambit 

Beneath the political noise of pulses and digital taxes, however, lies something substantial. For the first time in the bilateral relationship, the United States has explicitly committed to treating India as a trusted technology partner rather than a potential security risk. The interim framework includes a pledge to “significantly increase trade in technology products, including Graphics Processing Units (GPUs) and other goods used in data centres” . 

This is not routine trade facilitation. Under the Biden administration, the United States had placed India under GPU export restrictions, lumping it into a regulatory category that capped the number of advanced chips Indian companies could import. The logic was national security; the effect was to throttle India’s artificial intelligence ambitions before they could take flight. India’s IndiaAI Mission had managed to install roughly 40,000 GPUs through great effort and expense, a fraction of what a single American technology company might deploy for a single large language model . 

The Trump administration discarded those restrictions. More importantly, it enshrined India’s exemption in a bilateral trade agreement. This is the quiet victory that has received far less attention than the $500 billion import headline. India does not manufacture its own advanced semiconductors. It cannot, at present, produce Nvidia-grade GPUs. Its entire AI future depends on its ability to import American silicon. The interim framework guarantees that ability, and it does so in a form that cannot be easily revoked by a future executive order . 

The commercial implications are staggering. India’s data centre market is currently valued at approximately $10 billion, with revenue of $1.2 billion in fiscal 2024. Industry projections suggest this could grow to $100 billion in bilateral electronics trade alone. Google has already announced a $15 billion data centre partnership with the Adani Group. Microsoft has pledged $17.5 billion, primarily for AI infrastructure. Amazon has committed $35 billion over five years, though the precise allocation to data centres remains unspecified . 

Union IT Minister Ashwini Vaishnaw has gone further, suggesting that the combination of GPU access and the newly announced tax holiday—which exempts foreign data centre operators from corporate taxation until 2047—could spur investments of up to $200 billion. This is not government hyperbole; it is arithmetic. India currently pays GPU-as-a-service pricing nearly 40 percent higher than Singapore or the UAE, primarily because import duties of 20 to 28 percent on enterprise GPU servers eroded competitiveness. The interim framework signals duty rationalization. The industry estimates that even modest tariff reductions could lower data centre construction costs by approximately 14 percent, unlocking capital that was previously parked on the sidelines . 

This is where the deal stops being about trade balances and starts being about industrial transformation. India does not merely want to consume American chips; it wants to build the infrastructure that will allow it to eventually design its own. The interim framework includes provisions for discussing standards and conformity assessment procedures, technical language that conceals a more ambitious goal: the mutual recognition of American and Indian technology certification regimes. If successful, this would allow Indian-manufactured IT hardware to enter the United States without redundant testing, and American-designed components to enter India without licensing delays . 

The contrast with China is deliberate and explicit. The United States maintains a near-total prohibition on advanced GPU exports to Chinese entities. India, by contrast, has been granted open access, subject only to standard commercial terms. This is the strategic realignment that pulses and digital taxes obscure: Washington has decided that Beijing is the adversary and New Delhi is the hedge. That decision, once encoded in a trade agreement, is difficult to reverse. It is also, from India’s perspective, the single most valuable concession in the entire framework . 

The $500 Billion Question: Arithmetic or Aspiration 

And yet, for all the genuine progress on technology, the interim framework remains haunted by a number that few economists believe is plausible: $500 billion. 

This is the sum that India has stated it “intends” to purchase from the United States over five years. The linguistic retreat from “commits” to “intends” was necessary for domestic political consumption, but it did not change the underlying expectation. The Trump administration interprets “intends” as “will.” The Modi administration interprets “intends” as “aspires to, subject to market conditions and commercial viability.” These are not the same thing . 

Consider the arithmetic. Bilateral goods trade between India and the United States stood at approximately $132 billion in fiscal 2025, with India enjoying a surplus of roughly $41 billion. To reach $100 billion in annual American exports to India—the implied run rate of a $500 billion five-year target—India would need to more than double its current import volume from the United States, every year, for half a decade. This would require Indian companies to deliberately favor American suppliers over cheaper or more efficient alternatives, a form of administered trade that India has historically resisted and that its private sector is unlikely to embrace without substantial government coercion . 

The skepticism among economists is near-universal. Madhavi Arora of Emkay Global described the target as “more aspirational than realistic.” Biswajit Dhar, an independent trade expert, warned that $100 billion in additional annual imports “will completely upset India’s trade balance,” converting a durable surplus into a structural deficit. Dhiraj Nim of ANZ observed that markets have priced this skepticism, staging only a brief relief rally after the tariff announcement before settling back into wariness . 

The government has attempted to fill the $500 billion vessel with specific cargo. Boeing aircraft are frequently mentioned; Air India’s massive fleet expansion will indeed involve substantial purchases from the American manufacturer. Energy products—crude oil, liquefied natural gas, coking coal—are another plausible category, particularly if India accelerates its stated (and American-demanded) reduction in Russian energy imports. Information and communication technology goods, including the GPUs discussed above, will also contribute . 

But these are not增量 purchases. Air India would have bought Boeing planes regardless of the trade framework; the question is whether it buys more Boeing and less Airbus than commercial considerations would dictate. Indian refiners would have imported American crude if the price were right; the question is whether they are now expected to pay a premium for geopolitical alignment. The distinction between “intends” and “commits” matters precisely because it preserves, on paper, the principle of commercial autonomy while acknowledging, in practice, that Washington expects results . 

Shashi Tharoor, in a Lok Sabha debate that was notable for its precision, articulated the opposition case with unusual clarity. “No major economy has ever neutralised its own trade leverage in this manner,” he argued. “This is not strategic balancing; it is economic pre-emption.” His characterization of the deal as a “pre-committed purchase agreement” rather than a genuine reciprocal trade arrangement captures the unease that extends well beyond the Congress party. India has, in effect, accepted an obligation to reduce its trade surplus without securing corresponding commitments on services liberalization, visa access, or data flow protections. The asymmetry is difficult to defend on purely commercial grounds . 

A Deal Written in Pencil, Not Ink 

The quiet deletion of pulses and digital services taxes, the semantic retreat from “commits” to “intends,” the unresolved tension between GPU access and import targets—these are not bugs in the interim framework. They are features. They reveal that this is not a final agreement but a scaffolding on which a more complete arrangement might eventually be constructed. 

The Terms of Reference for the broader Bilateral Trade Agreement, expected by mid-March, list more than a dozen unresolved categories: non-tariff barriers, intellectual property, government procurement, labour standards, environmental provisions, and the treatment of state-owned enterprises. Each of these categories contains multiple points of genuine disagreement. India is unlikely to accept binding commitments on digital trade that would constrain its regulatory sovereignty; the United States is unlikely to accept Indian data localization requirements that disadvantage American cloud service providers. India wants H-1B visa expansion; the United States wants agricultural market access that extends beyond tree nuts and soybean oil. These are not easily reconciled . 

What the interim framework demonstrates, however, is that both capitals have concluded that continued negotiation is preferable to continued confrontation. The Trump administration inherited a bilateral trade relationship characterized by high tariffs, mutual recrimination, and a series of unresolved WTO disputes. The Modi administration faced the prospect of 25 percent retaliatory tariffs on Indian textiles, leather goods, and engineering products, a direct threat to the employment of more than two million workers in the electronics sector alone . 

The 18 percent reciprocal tariff—down from 50 percent, but still significantly higher than the zero-to-single-digit rates applied to most American trading partners—represents a compromise that neither side loves but both can tolerate. India retains meaningful tariff protection on politically sensitive sectors. The United States obtains a clear pathway to increased market share in India’s most dynamic import categories. Both sides avoid the reputational damage of a failed negotiation . 

The more revealing development, however, is what happened after the fact sheet was corrected. There was no diplomatic protest from Washington, no angry statement from the USTR, no threat to withdraw the tariff reduction. The White House simply updated the document and moved on. This suggests that the original overstatements were either a negotiating tactic that failed or a bureaucratic error that was swiftly contained. In either case, the absence of recrimination indicates that both governments understand the provisional nature of their work . 

The pulses will not be the last contested item to appear in a document and then disappear. The $500 billion figure will not be the last ambitious target that strains credulity. The interim framework is written in pencil, not ink, and both capitals are still editing. What matters is that they are still in the same room, still trading drafts, still treating each other as indispensable partners in a global economy that is rapidly dividing into hostile blocs. 

India will not become a zero-tariff economy overnight. The United States will not abandon its demand for reciprocal market access. The $500 billion gap between aspiration and arithmetic will not be closed by goodwill alone. But for the first time in years, the trajectory is positive. Tariffs are falling, not rising. Technology is flowing, not embargoed. Negotiations are continuing, not suspended. 

The quiet edit was not a defeat for the United States or a victory for India. It was a reminder that diplomacy, even at the highest levels, is an iterative process of claim and counter-claim, assertion and correction. The deal that exists today is not the deal that Washington originally described. But it is also not the deal that New Delhi originally feared. It is, instead, a careful compromise—provisional, imperfect, and still very much in progress.