Summary

  • GAO's July 2026 Technology Modernization Fund (TMF) report shows 24 projects expecting about $1.06 billion in combined savings using figures as of June 2025. In Figure 2, the Department of Veterans Affairs (VA) Veteran Identity Modernization project contributes $210.36 million to that dated total.
  • A footnote records a June 2026 VA revision: $35 million expected over fiscal years 2025–2028. VA attributed the reduction to identity-verification licenses it retained and kept paying for after omitting them from the previous cost estimate.
  • The official TMF page retrieved on November 3, 2026 describes the VA project as inactive, with all investment funds transferred and 67% of expected repayment recorded. Those are project and fund-flow statuses. A versioned savings record would join the older chart, the revised forecast, the cost assumption and later realized savings.

The million-dollar correction under the picture

The GAO report published July 23, 2026 offers a familiar public-finance picture: a large bar for future savings from upgrading old federal technology. Its June 2025 data showed 24 TMF projects expecting about $1.06 billion in total savings. Three projects supplied about $849.3 million of the forecast. The Department of Agriculture's Network Modernization led with $578.76 million; VA's Veteran Identity Modernization supplied $210.36 million; an Office of Personnel Management retirement-system project contributed $60.2 million.

Look under the VA bar. In comments on the draft report in June 2026, VA officials said they had revised the identity project's expected savings to $35 million over fiscal years 2025 through 2028. They explained that the project kept identity-verification licenses and kept paying for them. The earlier estimate had left those licenses out. The revision is large enough to change how a reader sees the project, and it is written into the same official report that carries the older figure.

GAO labeled the chart as of June 2025 and preserved that data snapshot. The footnote carries a later agency statement. Both dates matter. Treating the chart as an undated current promise discards the correction; treating the later $35 million as an automatic substitute in the whole portfolio sum would mix forecast versions and periods that the report has not reconciled. The public needs a joined record showing which project estimate was used in each aggregate, when it changed, why, and what comparison period each version covers.

The licenses survived the modernization

The VA project offers a clean illustration of how a savings forecast can lose its engine. It was meant to simplify and strengthen access to veterans' services. According to the official TMF project page, VA retired the My HealtheVet user ID and password in March 2025 and ended support for DS Logon at VA in November 2025. Users now face two modern credential options instead of four. That is a tangible service change.

The financial assumption was different. VA's 2026 comment to GAO said it retained identity-verification licenses it still had to pay for. A retired login method and a retained verification bill can coexist. That is the conflict behind the estimate revision: the user-facing transformation happened while a cost the savings model expected to shed remained in place. A project can deliver access improvements and still produce less cash saving than its first spreadsheet promised.

The TMF page currently lists the project as inactive, with 100% of TMF investment transferred, 67% repayment status, no reported schedule delay or cost overrun, and total TMF investment of $10,534,628. These status fields answer useful questions about project execution and fund flows. None supplies a realized-savings figure on the page. The gap between project-status reporting and forecast-history reporting is precisely why the two GAO numbers need a durable link rather than a casual replacement.

The rest of the billion has its own clock

GAO's broader TMF findings need equal care with dates. The board had invested about $1.03 billion in 68 unclassified projects as of June 2025. Of those, 24 expected savings; 37 pursued other value, including security and service improvements; seven had been cancelled before that cutoff and no longer expected savings. Eleven of the 24 forecast-bearing projects had realized about $13.5 million by then, while 13 had not yet begun to realize savings.

The $13.5 million is an early realized amount, not a verdict on the $1.06 billion forecast. GAO said 21 projects carrying roughly $1.04 billion of the expected savings planned to finish realizing them in fiscal year 2027 or later. Twelve of those expected to finish in 2030 or later. A responsible account needs to show the calendar alongside the total. Waiting for a future year is part of the original forecast; retaining a license cost after the estimate was built is a change in the forecast's economics.

The completed-project subset tells a different story again. GAO assessed six projects with at least a full fiscal year of savings after completion. Two met or were on track to meet their targets within the auditor's 10% test; four did not meet or were not on track. Officials cited reduced project scope and higher migration costs for some shortfalls. The denominator is six, not the whole portfolio. It shows why assumptions deserve maintenance while the projects are still underway.

Put the revision where the decision is made

The Technology Modernization Board, GSA, OMB and recipient agencies each have a role in approving or administering the fund. Agencies supply savings estimates; GSA publishes project information; GAO audits the record for Congress. That division allows the original forecast and the later correction to live in different places. The chart is easy to repeat. The note requires a second read. When public money is justified partly by future savings, that layout matters.

The missing public object is a forecast history for each project: original estimate and as-of date, revised estimate and period, changed cost or scope assumption, realized savings to date, and the version included in a portfolio aggregate. In VA's case, the first line would carry $210.36 million in the June 2025 series; the second would carry the June 2026 $35 million revision and its FY2025–2028 horizon, with retained license payments named as the reason. A current aggregate would need comparable versions from the other projects before claiming a new total.

The next official update that puts those pieces on one record would change the judgment. For now, GAO has supplied the facts but left the reader to stitch the old bar to the new note. A forecast can survive a footnote. It should not outrun it.

Sources