The core of her agenda for antitrust is not about being anti-merger or anti-bigness. Rather, her project is mainly about creating the conditions for pluralistic, bottom-up and more "permissionless" markets. Markets where founders can build successful businesses without necessarily having to rely on large 'gatekeeper' firms as partners. To the extent that large firms are required, Khan's vision is that they shouldn't be able to use their economic power to bargain unfairly or exploitatively with smaller firms.
Part of her agenda is strengthening the merger review process. Merger review is intrinsically hard, because authorities essentially have to try and figure out what the effect will be if the merger is allowed or denied. There is no crystal ball for this, and there are rarely 'right' answers. Yet the merger regime has been very favourable to large firms over the past few decades, and there's a reasonable consensus in the antitrust community that it should be strengthened. A tighter merger process might make startup exits through acquisition less common, but it should also make it easier for startups to grow organically into large companies. In terms of creating a bottom-up and more pluralistic Silicon Valley, that seems like a win to me.
(Source: I'm doing a PhD in Antitrust/Competition Law)